In optoelectronic devices, light rays are absorbed and generate charge carriers within the device. These charge carriers typically are desired to be generated within a particular light ray absorption region, which can be defined by a depth within the device, such that they can be collected near the surface of the device.
Charge carriers generated deeper than the depth of this region can be thought of us undesirable noise. Conventional approaches to dealing with these charge carriers often relate to transporting them to the surface by extended electric fields or annihilating them by the targeted introduction of recombination centers. The former is not suitable in all situations, such as those with regions that must remain free of electrical fields for physical reasons, and is also limited by available voltage, while the latter reduces internal quantum efficiency and can be technically difficult to realize at very high doping atom densities. Reduced quantum efficiency in turn can affect devices dimensions, and increased complexity and technological challenges can increase costs, which are undesired.
A variety of optical sensors use infrared light for recognition of objects and distances. It is known that with increasing wavelength, the depth of penetration of light into a silicon structure increases. With the increasing depth of penetration, for time critical applications, the delay of slow diffusion currents from the greater depth to which the infrared light penetrates to the surface of the devices is slow. This can create time delays that become significant in view of the greater time required for the slow diffusion currents to reach the surface. To address these issues, for infrared wavelengths in the range of 900 nm, the prior art commonly uses electrical confinement such as built in fields. However, these electrical confinement efforts can become costly and can increase production costs to a level that is economically unfeasible.
In a wide range of applications, infrared light is the light signal of choice because it is invisible to human beings. This permits measurements to be made without the awareness of human observers. In the prior art, such as proximity sensors or time of flight sensors, losses of quantum efficiency are accepted in most cases. For time critical applications, losses in quantum efficiency are sometimes even increased by implementation of recombinative regions in deeper areas of the silicon structure in an effort to eliminate disturbing background current that arises from slowly diffusing charge carriers. A graph depicting the increased penetration depth with longer wavelength is depicted in FIG. 7.
In the field of photovoltaics, in the prior art, a variety of so-called “light trapping” techniques have been disclosed. In the photovoltaic prior art, a variety of structures have been investigated for improving the light performance including inverted pyramids, dice-grooving of multi-crystalline material or perpendicularly oriented grooves on the front and the back surface of the photovoltaic device. However, these techniques are unsuitable for use in CMOS integrated circuit applications because of the structures themselves. An effective photovoltaic uses a front and a rear surface that directly covers the active absorption region of the optical electronic silicon device. In standard CMOS technologies, the rear side of the structure is defined by back thinning in back end processing. Accordingly, dedicated structures for light trapping cannot be defined during a front end fabrication sequence.
Many optical electronic silicon devices benefit from having a small pixel size. High internal quantum efficiency is important to achieving small pixels sizes which require small device dimensions. In particular, in time critical applications, it is important to define a small absorption region to reduce noise components that are produced by charge carriers that are generated deep in the volume of the semi-conductor. As discussed above, this can be difficult because silicon as an indirect semi-conductor absorbs infrared wavelength spectral light components only weakly. Because of this, the penetration depth for light at wavelengths of between 800 and 900 nm, is approximately 12 micrometers to 32 micrometers. In situations where it is beneficial for photo generated charge carriers to be collected near the surface of the semi-conductor, this property of the material can be problematic for time critical device operation. According to the prior art, to address this problem, charge carriers that are generated deep in the volume of the silicon semi-conductor have been transported to the surface by the application of extended electric fields. These electric fields are built up in lightly doped zones and are limited by the available voltage. Accordingly to the prior art, regions of the semi-conductor that need to remain free of a field for physical reasons are designed such that the electron hold pairs generated therein were annihilated by the targeted introduction of recombination centers. This approach leads to reduced quantum efficiency and is technologically difficult to implement at very high impurity atom densities. Under the circumstances, the recombination active zone is overgrown with a lightly doped epitaxial layer, which is readily susceptible to being dislocated on such a support.